University of Illinois at Urbana-Champaign

NIH Resource for Macromolecular Modeling and Bioinformatics

Beckman Institute
Computational Biophysics Workshop

Aquaporins

Fatemeh Khalili
VMD Developers: Elizabeth Villa
John Stone Yi Wang
Dan Wright Kwok Yan Chan
John Eargle Emad Tajkhorshid
Brijeet Dhaliwal
Zan Luthey-Schulten

July 2009
 2

A current version of this tutorial is available at

http://www.ks.uiuc.edu/Training/Tutorials/
CONTENTS 3

Contents
1 Introduction to Aquaporin Structure 7
1.1 Structure and Function of Aquaporins . . . . . . . . . . . . . . . 7
1.2 Loading other aquaporin structures . . . . . . . . . . . . . . . . . 13

2 Structural Alignment of Aquaporins 15

2.1 Starting Multiple Sequence Alignment . . . . . . . . . . . . . . . 15
2.2 Setting parameters for Multiple Structure Alignments . . . . . . 15
2.3 Aligning the molecules . . . . . . . . . . . . . . . . . . . . . . . . 16

3 Comparing Protein Sequence and Structure 19

3.1 Protein Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Protein Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

4 Residue Selection 23
4.1 Structure conservation . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2 Sequence conservation . . . . . . . . . . . . . . . . . . . . . . . . 26

5 Investigating Structural Alignment 28

5.1 RMSD per Residue . . . . . . . . . . . . . . . . . . . . . . . . . . 28

6 Examining the Aquaporin Tetramer 31

6.1 Loading Tetramer . . . . . . . . . . . . . . . . . . . . . . . . . . 31
6.2 Examining channel lining and tetrameric contacts . . . . . . . . . 32

7 Phylogenetic Tree 34

8 Evolutionary Profile of AQPs 36

8.1 Configure BLAST for Multiseq . . . . . . . . . . . . . . . . . . . 36
8.2 Load Structures for AQPs in all three domains of life . . . . . . . 37
8.3 Load Sequences for AQPs in all three domains of life . . . . . . . 38
8.4 Align Sequences Using a Structural Profile . . . . . . . . . . . . . 40
8.5 Construct an Evolutionary Profile for AQPs . . . . . . . . . . . . 40
CONTENTS 4

Introduction
The Multiple Sequence Alignment extension to VMD
The Multiple Sequence Alignment (Multiseq) extension to VMD links protein
structures to protein sequences and allows you to compare proteins in terms of
structure and sequence.
Multiseq is designed to study protein mechanism through available genetic in-
formation, such as evolutionary conservation, offering biomedical researchers a
new tool to examine protein structure and function.
This tutorial can be used by both new and previous users of VMD. However,
it is recommended that new users go through the “VMD Molecular Graphics”
tutorial, in order to gain further knowledge about the overall program.1
The present tutorial has been designed specifically for VMD with Multiseq and
should take about an hour to complete in its entirety.

Aquaporins
We will use the family of aquaporins for a case study in the applications of the
Multiple Sequence Alignment tool. Aquaporins (AQPs) are membrane chan-
nel proteins found in a wide range of organisms, from archaea and bacteria to
plants and animals. AQPs facilitate the rapid transport of water across cellular
membranes and are of fundamental importance to the control of cell volume and
transcellular water traffic.
In humans, there are at least 11 different aquaporin types (see cover). The
kidney alone contains AQP1, AQP2 and AQP3. These proteins are responsible
for filtering hundreds of liter of water per day in the human body. In addition
to water, a subfamily of AQPs, named aquaglyceroporins, also selectively trans-
port small molecules such as glycerol.

Here we use Multiseq to conduct a comparative study of the structure and

sequence of four aquaporins from different species: human AQP1, bovine AQP1,
AqpZ from E.coli, and GlpF (E.coli glycerol facilitator). The aquaporin AqpM
from Archaea will be introduced in the last section.

1 URL: http://www.ks.uiuc.edu/Training/Tutorials/
CONTENTS 5

Getting Started
Multiseq, in its current release, is operable on the following platforms:
• Macintosh OS X

• Solaris
• Linux
• Windows
This tutorial requires certain files, located in

/Desktop/Workshop/aqp-tutorial/aqp-tutorial-files/

The file structure of this directory appears in the following figure:

Needed Software
VMD, containing the MultiSeq plugin (version of VMD 1.8.7 beta 6 and later),
can be obtained from http://www.ks.uiuc.edu/Research/vmd. Once VMD is
installed, you can start the tutorial by starting VMD. To start VMD:

• Mac OS X: Double click on the VMD application icon in the Applications

directory.
• Linux and SUN: Type vmd in a terminal window.
• Windows: Select Start → Programs → VMD.
CONTENTS 6

Getting Help
We have set up a mailing list, Tutorials-L, which will act as a forum for dis-
cussions and questions about the tutorials offered by the Theoretical and Com-
putational Biophysics Group (http://www.ks.uiuc.edu/Training/Tutorials/). If
you have questions, comments, etc, please suscribe to the list. Instructions can
be found at http://www.ks.uiuc.edu/Training/Tutorials/mailing list.
1 INTRODUCTION TO AQUAPORIN STRUCTURE 7

1 Introduction to Aquaporin Structure

In this unit you will use the conventional molecular graphics tools of VMD to
become familiar with the key structural features of aquaporins, and how these
are related to aquaporin function.

1.1 Structure and Function of Aquaporins

You will first consider bovine aquaporin (PDB code 1J4N).

In order to learn about the structural features of aquaporins, you will create
several graphical representations of the molecule. Here, we will only guide you
through the steps to create the necessary representations. If you want to know
more about graphical representations, please look at the main VMD tutorial2 .

Using a VMD saved state. If you are familiar with VMD and don’t
want to go through all the steps of creating representations, you can
use the VMD saved state aquaporin.vmd located in your tutorial
directory to load the molecule with all the necessary representation.
However, we recommend that you take the time and go through
the suggested steps. If you choose to use the saved state, you still
need to read through this section to learn about important structural
features of aquaporins.
We also note that in case you need to interrupt the tutorial, you
can save the current state of the VMD session (see VMD tutorial)
and resume your study at a later time.

1 Choose the File → New Molecule... menu item in the VMD Main window.
Another window, the Molecule File Browser, will appear on your screen.

2 Use the Browse... button to find the file 1j4n.pdb in the directory
aqp-tutorial-files → PDB in the tutorial directory. Note that when
you select the file, you will be back in the Molecule File Browser window.
In order to actually load the file you have to press Load. Do not forget to
do this!

Getting a PDB file over the Internet. In case your computer is

connected to the Internet, you can use VMD to get the file 1j4n.pdb
automatically from the Protein Data Bank3 (see VMD tutorial).

Now, the aquaporin is shown on your screen in the OpenGL Display window.
You may close the Molecule File Browser window at any time.
2 URL: http://www.ks.uiuc.edu/Training/Tutorials/
3 URL: http://www.pdb.org
1 INTRODUCTION TO AQUAPORIN STRUCTURE 8

3 Choose the Graphics → Representations... menu item. A window called

Graphical Representations will appear and you will see the current graph-
ical representation used to display your molecule highlighted in light green.

4 In the Draw Style tab you can change the style and color of the represen-
tation. In the Selected Atoms text entry of the Graphical Representations
window, delete the word all. In the place of all, type protein and press
the Apply button in the bottom right-hand corner of the window, or hit
Enter or Return key on your keyboard. It is important that you do this
every time you type something in Selected Atoms. From Drawing Method,
choose the Tube menu item. The representation draws a tube along the
backbone of the protein (Fig. 1). Now, change the color of the molecule
by choosing the Molecule menu item from the Coloring Method menu.

Figure 1: Tube representation of aquaporin structure.

The tube representation you created will be used throughout the tutorial to
look at the structural alignment of aquaporins, but before you start using the
Multiseq program, you will create other representations in order to learn some
of the important structural features of aquaporins.

5 In the Graphical Representations window, click the Create Rep button.

This will create a new representation, identical to the one you had before.
This time, choose the drawing method NewCartoon. This drawing method
shows the secondary structure of aquaporins (Figure 2).

6 In the Graphical Representations window, double–click on the first rep-

resentation. This will make the text of the first representation red in
this menu, and will hide the corresponding representation in the OpenGL
Display window.
1 INTRODUCTION TO AQUAPORIN STRUCTURE 9

Figure 2: The NewCartoon representation of an aquaporin structure.

Secondary Structure of Aquaporins. Aquaporins are composed

of six transmembrane α-helices and two reentrant loops, formed by
a short helix and an extended polypeptide that returns to the surface
of the protein. The two reentrant loops, which are surrounded by the
six long helices, meet each other in the middle of the channel. These
loops line the channel and are of great functional importance.

7 Click on the OpenGL window and use your mouse to rotate the molecule
and look at the structure of aquaporins.
For the rest of this section, you will continue to look at the aquaporin structure
in Tube representation. You can always go back and look at the NewCartoon
representation, if necessary.
8 In the Graphical Representations window, double click on the first repre-
sentation. This will show it again in the OpenGL window. Double click
on the second representation (NewCartoon) to hide it.
Now look closer at the structural details of the aquaporin structural elements.
9 Create a new representation, clicking on the Create Rep button. This time,
you will focus on the helices of the reentrant loops mentioned above (see
box). In the Selected Atoms text entry of the Graphical Representations
window type resid 79 to 88 196 to 204. These residues correspond to
two helices of aquaporin that face each other in the middle of the channel.
Choose the coloring method ColorID → 1 and the drawing method Tube.
1 INTRODUCTION TO AQUAPORIN STRUCTURE 10

10 Repeat step 9. This time in Selected Atoms, type resid 74 to 78 191

to 195, and for the Coloring Method, select ColorID → 7. These residues
correspond to the extended polypeptide regions of the reentrant loops of
aquaporins (see box).

Now that you have localized the reentrant loops, the rest of the protein will
appear dim.

11 Click (once) on the first representation. In the Material menu, choose

Transparent. Your OpenGL window should look like Fig. 3.

Figure 3: The reentrant loops are a key structural feature of aquaporins. Each reentrant
loop is formed by a short helix, shown in red, and an extended polypeptide, shown in green.

Now that you can locate the reentrant loops, you will examine the NPA motif
within the reentrant loops.

NPA motif. The NPA motifs, as the name implies, are formed by
amino acids N – asparagine, P – proline, and A – alanine. They are
highly conserved signature motifs in all aquaporins. They stabilize
the reentrant loops through multiple hydrogen bonds, as shown in
figure 4.

12 Create a new representation, by clicking on the Create Rep button. Make

sure the new representation is not transparent. In the Selected Atoms text
entry of the Graphical Representations window, type resid 78 to 80
194 to 196. These residues correspond to the two NPA motifs present
in aquaporin. Choose the drawing method Licorice and coloring method
Type. Look at the NPA motifs, can you see how they would be hydrogen
bonded?
1 INTRODUCTION TO AQUAPORIN STRUCTURE 11

Figure 4: The NPA motifs in aquaporins.

13 Create another representation by clicking the Create Rep button. With the
same coloring and drawing methods of the previous selection, type resid
197 in the Selected Atoms text window. This will draw an important
arginine residue in aquaporins.

Figure 5: The conserved arginine provides H-bond sites to water molecules inside the chan-
nel.

Conserved arginine. This arginine is the only charged residue in

the channel, and is conserved in all aquaporins. This arginine lines
the pore and provides additional H-bond sites for water molecules
permeating the channel (See Fig. 5).
1 INTRODUCTION TO AQUAPORIN STRUCTURE 12

14 Repeat step 13, but this time type resname GLU in the Selected Atoms
text window. This will select all the glutamates in the protein. Notice
that there are only two glutamate side chains in the transmembrane region
that are in the helical core of the protein. If you look closely (c.f. Fig. 6)
you will see that each one of these glutamates sits behind a reentrant loop.
Glutamates. The glutamates serve to stabilize the structure of the
reentrant loops. They form direct hydrogen bonds with the loops
and thus hold them in their place (Fig. 6). These glutamates are
also highly conserved in the aquaporin family. It is interesting that
mutation of one of these glutamates results in cataracts in human
eyes!

Figure 6: The glutamates help stabilize the reentrant loops.

Finally, you are going to look into the aquaporin pore.

15 Create a representation. In the Selected Atoms text window type protein.

Set the drawing method to VDW (van der Waals). Each atom is now rep-
resented by a sphere. In this way you can view the volumetric distribution
of the protein.
16 Rotate the molecule so that you can see it from the top. Look down the
pore. Can you see the pore? How many water molecules you think fit
there?

Pore. The pore is so narrow (Fig. 7) that it can fit water molecules
only in single file. This feature is very important for the selectivity
of the channel.
1 INTRODUCTION TO AQUAPORIN STRUCTURE 13

Figure 7: The pore of aquaporins can only accomodate water molecules in single file.

17 Create a new representation. In the Selected Atoms text window type

water. Set the Drawing Method to VDW and the Coloring Method to
type. This will create a representation with all the water molecules the
crystallographers resolved in the crystal structure. You can now see that
the pore is filled with water molecules forming a single file.
Now that you have learned about the structural features of aquaporins, you will
load the other three aquaporins mentioned above and get ready to align them.
Before that, you should turn off all representations created so far, leaving only
the first.
18 In the Graphical Representations window, double click on each of the
graphical representations, except the first. This will make them appear
in light color in this menu, and will hide the corresponding views in the
OpenGL window. Select the first representation and set the material to
Opaque.

1.2 Loading other aquaporin structures

In this tutorial you will be comparing four aquaporin structures listed in Ta-
ble 1.2. You have already loaded the 1j4n molecule. Now, you will load the
other three aquaporins.

PDB code Description

1j4n Bovine AQP1
1fqy Human AQP1
1lda E. coli Glycerol Facilitator (GlpF)
1rc2 E. coli AqpZ

Table 1: Aquaporin structures

1 The remaining aquaporins, 1fqy, 1lda, 1rc2, need to be loaded into VMD
and have their molecule’s graphical representations individually changed
1 INTRODUCTION TO AQUAPORIN STRUCTURE 14

to Tube. To do this, you need to refer back to the previous section, Struc-
ture and Function of Aquaporins, and repeat steps 1 through 4. Make sure
that each PDB is loaded into a new molecule.

Figure 8: Aquaporin structures loaded into VMD.

You should now see the four molecules loaded in the OpenGL Display (Figure 8).
If you do not see all molecules, hit S on the keyboard while in the OpenGL
display. Then use your mouse to scale the molecules in the display, such that
you can see all four. Note that the structures are not aligned, but rather placed
according to the coordinates specified in their PDB files.

2 Close the Graphical Representations window and Molecule File Browser

window.

1rc2. The original 1rc2 file contains two copies of the aquaporin
molecule. This is the way the crystallographers reported this struc-
ture to the PDB since the crystal used for analysis contained two
independent copies of AqpZ. For this tutorial, we include only one
copy of the molecule in the provided pdb file.

You are now ready to start performing Multiple Sequence Alignment.

2 STRUCTURAL ALIGNMENT OF AQUAPORINS 15

2 Structural Alignment of Aquaporins

In this section, you will align the aquaporin molecules loaded in VMD using the
Multiseq program.

2.1 Starting Multiple Sequence Alignment

Now that you have loaded the four proteins in Table 1.2, you will use Mul-
tiseq program to align the proteins and analyze their structural and sequence
relationships.

Figure 9: Multiseq Program Window

1 Within the VMD main window, choose the Extensions menu.

2 In the Extensions menu select Analysis → MultiSeq.

This is the main Multiseq program window. The rest of the tutorial and exercises
will use features from this window, unless specified otherwise. You may be asked
to update some databases if this is the first time you use Multiseq. If this is the
case, simply click Yes and wait for Multiseq to start.
By default, Multiseq will align all four loaded molecules, unless you delete
the molecule(s) in the VMD Main window. However, note that some crystal
structures come with water and detergent molecules, which should not be used
to align the structures. In your MultiSeq window, you will find that all the
protein structures are categorized as VMD Protein Structures and other
molecules, e.g. water, are in the VMD Nucleic Structures category. Keep
the four protein structures and delete the two structures under VMD Nucleic
Structures by clicking them and press the delete button on your keyboard.

2.2 Setting parameters for Multiple Structure Alignments

Multiple Sequence Alignment uses the program STAMP to align protein molecules.
2 STRUCTURAL ALIGNMENT OF AQUAPORINS 16

STAMP. The program STAMP (Structural Alignment of Multiple

Proteins) is a tool for aligning protein sequences based on a three-
dimensional structure. The STAMP algorithm minimizes the Cα
distance between aligned residues of each molecule by applying glob-
ally optimal rigid-body rotations and translations. Also, note that
you can perform alignments on molecules that are structurally sim-
ilar. If you try to align proteins that have no common structures,
STAMP will have no means to align them. If you would like fur-
ther information about how the alignment occurs, please refer to
the STAMP manual at:
http://www.rfcgr.mrc.ac.uk/Registered/Help/stamp/stamp.html

Figure 10: Stamp Alignment Options

Before you align the molecules, you may want to change the default parameters
needed by STAMP.

1 Choose the Tools → Stamp Structural Alignment tool. The window shows
several parameters that can be set (Figure 10). For this tutorial, you will
use the default parameters.

You can look at the STAMP user guide (see box) for information on how to
optimize the STAMP parameters to obtain a better alignment.

2.3 Aligning the molecules

Now that you have opened the MultiSeq window and made sure the STAMP
parameters are correct, you can align the molecules you loaded into VMD.

1 In the Stamp Alignment Options window, choose Align the following: All
Structures and go to the bottom of the menu and select OK.
2 STRUCTURAL ALIGNMENT OF AQUAPORINS 17

Figure 11: Structural alignment of four aquaporin molecules.

The molecules have been aligned. You can see this both in the OpenGL window
(Fig. 11) and in the MultiSeq program window (Fig. 12). Take some time to
look at the alignment. Note that your alignment may not immediately resemble
Fig. 11. This is because MultiSeq displays the protein in “NewCartoon” repre-
sentation by default. To compare your alignment result with Fig. 11, change
the representations for each protein to “Tube” in your Graphical Representations
window. You could also change the color for each protein.

2 Click on the OpenGL window (Fig. 11). Move your mouse around, rotat-
ing the molecules. Do you think this is a good alignment? Look at the
top of the molecules: Can you see the pore?

Figure 12: Sequence Alignment of four aquaporin molecules.

3 Now, click on the Multiseq program window (Fig. 12). Move the left–to–
right scroll at the bottom of the window, and take a look at the residues in
the Sequence Display. Observe that there are gaps, represented by dashes,
in the sequences when they are aligned.
2 STRUCTURAL ALIGNMENT OF AQUAPORINS 18

In the next section, you will start using Multiple Sequence Alignment features
for the alignment you just made.
3 COMPARING PROTEIN SEQUENCE AND STRUCTURE 19

3 Comparing Protein Sequence and Structure

In this unit you will study the structure and sequence conservation of the differ-
ent aquaporins you have aligned in the Multiseq window. Conservation is a term
we will use frequently in exercises from this point forward. Within the context of
protein analysis, conservation refers to high levels of similarity contingent on the
employed metric. Structure conservation occurs when the structural aspects (e.g.
α helix, β sheets), of the aligned proteins are highly similar. Likewise, sequence
conservation happens when at certain points of the aligned protein molecules the
amino acid of each molecule is the same.

3.1 Protein Structure

In order to better understand the structural conservation within the aligned
molecules, Multiseq provides tools that allow you to color the molecules accord-
ing to their Q value (Qres), a measure of structural conservation.

Q per residue. Q is a parameter that indicates structural identity.

Q accounts for the fraction of similar native contacts between the
aligned residues in two proteins4 . Q=1 implies that structures are
identical. When Q has a low score (0.1-0.3), structures are not
aligned well, i.e., only a small fraction of the Cα atoms superimpose.
Q per residue is the contribution from each residue to the overall Q
value of aligned structures.

You will now color the molecules according to the value of Q per residue obtained
in the alignment.
1 In the Multiseq program window, choose the View → Coloring → Qres
(Fig. 13).

2 Look at the OpenGL window to see the impact this selection has made
on the coloring of the aligned molecules (Figure 14).

Rotate the molecule to see how much of it has turned blue. Notice that the
transmembrane helices of the aligned molecules have turned blue. The blue
areas indicate that the molecules are structurally conserved at those points. If
there is no correspondence in structural proximities at these points, the points
appear red. Observe how the structurally least similar segments tend to be on
the periphery of the aligned molecule. Note that the loops tend to be red, while
the helices are blue.
4 Eastwood, M.P., C. Hardin, Z. Luthey-Schulten, and P.G. Wolynes. “Evaluating the

protein structure-prediction schemes using energy landscape theory.” IBM J. Res. Dev. 45:
475-497, 2001. URL: http://www.research.ibm.com/journal/rd/453/eastwood.pdf
3 COMPARING PROTEIN SEQUENCE AND STRUCTURE 20

Figure 13: Molecule Coloring

Figure 14: Molecules colored by Qres value

3.2 Protein Sequence

You have examined the structural similarity between the molecules. Now you
will look at the sequence conservation. Multiseq has a feature to color the
molecules according to how much the sequence is conserved within the aligned
molecules. This tool, Sequence identity, colors each amino acid according to
the degree of conservation within the alignment: blue means highly conserved,
whereas red means very low or no conservation.

1 Choose View → Coloring → Sequence Identity.

Before you look at the viewer window, can you anticipate what will happen to
the coloring of the molecules? Will the molecules still be blue in the trans-
membrane region, as they were when Qres was used to determine structure
conservation?
3 COMPARING PROTEIN SEQUENCE AND STRUCTURE 21

2 Now take a look at the viewer window. As you can see (Fig. 15), a fair por-
tion of the molecules has turned red, indicating less sequence conservation
than structural similarity.

Figure 15: Molecules colored by sequence identity

3 Note that the residues at the site where the two short helices interact are
blue. Look at the molecules from the top (c.f. Fig. 16). Do you notice the
blue residues tend to be on the inside of the pore?

Figure 16: Conserved residues (shown in blue) are located inside the pore.

Note that the coloring of the molecules using Sequence Identity indicates
that the sequence conservation is much less in comparison to the structural
3 COMPARING PROTEIN SEQUENCE AND STRUCTURE 22

conservation. Sometimes structures from the same family have less than 10%
sequence identity, yet are structurally similar. A well known example is the
protein myoglobin that we recommend for self-study.
To examine the relationship between sequence and structure in more detail, in
the next section you will use the Select Residues feature.
4 RESIDUE SELECTION 23

4 Residue Selection
In this section, you will use the Select Residues tool for finding structural fea-
tures of aquaporins that have been conserved throughout species and are crucial
for their function.

The Select Residues features analysis of structural similarity and sequence con-
servation. It makes use of different measures to highlight residues in the Se-
quence Display and Structure Display simultaneously. It allows you to examine
the conservation and similarity on a per residue basis, using two different mea-
sures, Qres value measuring structure similarity and Sequence Identity measuring
sequence conservation.

4.1 Structure conservation

You can use Multiseq to look at the residues that are structurally similar above
or below a certain value of Qres.

Figure 17: Residue Selection

1 Choose the Search → Select Residues tool. A side menu will appear
(Fig. 17). Within the menu, you can select residues based on their Se-
quence Identity or Qres.
2 In box following Qres, select ≥ and key in 0.5 for the value.
3 Click on the Select button.
Note the changes in the display of the Open GL (structure) Display window
(Fig. 18, left) and in the Sequence Display of the Multiseq program window
(Fig. 19).
4 RESIDUE SELECTION 24

4 Look in the Sequence Display (Fig. 19). Since you selected the Qres value
to be greater or equal to 0.5, most of the structural elements of the aligned
molecules will be shown as conserved (similar), and will be displayed in
yellow.

Figure 18: View of the Residue Selection for Qres≥0.5 in the Open GL Display window.
Highlight style set to Licorice (left) and NewCartoon (right)

5 Now, look at the OpenGL Display window (Fig. 18, left). In this case also
most parts of the molecules are highlighted, in particular, the transmem-
brane segments.

Figure 19: View of the Residue Selection for Q≥0.5 in the Sequence Display window.

6 Due to the high level of conservation, the Licorice highlighting style is

not very informative. To improve the view, chose the View → Highlight
Style menu item, and click on the New Cartoon style (Fig. 18, right). You
can recognize immediately that high Qres values indicating conservation of
neighborhood relationships for amino acids are seen in the transmembrane
region throughout.
4 RESIDUE SELECTION 25

Now you will investigate what happens for higher values of Qres.

7 In the Select Residues window, select ≥ after the Qres and key in 0.75.
Click on the Select button.

8 Look at the OpenGL Display window (Fig. 20). Notice that most of the
structurally similar regions are still in the transmembrane region, but the
highlighted regions are more confined than before.

Figure 20: View of the Residue Selection for Qres≥0.75 in the OpenGL Display window

9 Look in the Sequence Display window (Fig. 21). You can still distinguish
blocks of structurally similar residues, which correspond to the helices.
They are not as well defined as before, but the structural similarity is still
noticeable when probing for high values of Qres.

Figure 21: View of the Residue Selection for Qres≥0.75 in the Sequence Display window
4 RESIDUE SELECTION 26

10 Repeat steps 7 to 9 and set the Qres to ≥0.9. You will notice that almost
all the structure conservation is lost.

What does this mean? The aligned aquaporins have a moderate to high struc-
ture conservation. Structurally, the most similar areas are the transmembrane
helices.

4.2 Sequence conservation

You can also use Select Residues to look at the residues that are conserved on
the sequence level.
1 Go back to the main Multiseq program window and select Search → Select
Residues.

Figure 22: View of the Residue Selection by Sequence Identity in the OpenGL Display
window.

2 In the Select Residues window, following Sequence Identity select ≥.

3 You can then key in a value between 0 and 100. Try 70, which will select
residues that were assigned a sequence identity measure of 70% or more.

4 Click on the Select button.

5 Look at the Sequence Display in the Multiseq program window. You will
notice very few residues are selected. If you look carefuly, you will locate
some of the residues that correspond to key structural features of aqua-
porins, for which you made representations in section 1 of this tutorial.
Search for the NPA motifs, highlighted in yellow (Fig. 23).
4 RESIDUE SELECTION 27

Figure 23: View of the Residue Selection by Sequence Identity in the Sequence Display
window

6 In the OpenGL window, look at the areas highlighted in yellow. Notice

that many residues in the reentrant loops (Fig. 3) conserved their sequence
over the course of evolution. This is a clear indication that the reentrant
loops are functionally significant.

The conserved amino acids in the sequence are mostly located in the pore, and in
the reentrant loops. You will now compare these residues to the representations
you created in section 1 of this tutorial, which correspond to residues that are
relevant to the function of aquaporins.

7 In the Multiseq program window, select View → Highlight Style → Licorice.

8 In the VMD Main window, select Graphics → Representations....

9 In the Graphical Representations window, in the Selected Molecules pull-

down menu, select the molecule 1j4n.

10 In the Graphical Representations window, double click on the represen-

tations you created for the NPA motif, the glutamates and the arginine.
What is the sequence identity for these amino acids? Why do you think
the NPA motif, the lysine and the glutamates are conserved throughout
species?

When you are done examining the NPA motif, close the Graphical Representa-
tions window to proceed to the next section.
5 INVESTIGATING STRUCTURAL ALIGNMENT 28

5 Investigating Structural Alignment

In this section you will use RMSD values between molecules as a measure to
see how close structures of molecules are after alignment.

5.1 RMSD per Residue

RMSD per residue values compare two proteins and show how well they align.
Here we use RMSD values that compare solely the Cα atoms of the proteins.

RMSD values. The Root Mean Squared Deviation (RMSD) is a

numerical measure of the difference between two structures. It is
defined as :
s
PNatoms
i=1
(ri (1) − ri (2))2
RM SD = (1)
Natoms

where Natoms is the number of atoms whose positions are being

compared, and ri (1) and ri (2) are the position of atom i in each
molecule.

1 In the Multiseq program window choose Tools → Plot Data. A window

entitled Data Plotting will appear on your screen. (Fig. 24)

2 Select the All sequences option.

3 Choose the RMSD checkbox, for the Per-residue data to plot.

Figure 24: Data Plotting window.

5 INVESTIGATING STRUCTURAL ALIGNMENT 29

4 Click on the OK button to produce a plot of the RMSD values between the
first molecule in the Sequence Alignment window and any of the aligned
structures (Fig. 25). The RMSD values between each pair of the aligned
residues are shown along the sequence of the protein, exhibiting regions of
close alignment (small RMSD values) and of poor alignment (large RMSD
values).

Figure 25: RMSD values between bovine AQP1 (pdb code:1j4n) and human AQP1 (1f1y),
E.Coli GlpF (1lda), and E.Coli AqpZ (1rc2)

5 Go back to the Multiseq program window. Compare the RMSD graph

with the sequence alignment, and identify the residues with large RMSD
values. What structural elements do they correspond to? As you will see
later, they correspond to the loops.

Figure 26: Multiseq program window. Residues corresponding to the first peak in the RMSD
graph are highlighted in yellow.
5 INVESTIGATING STRUCTURAL ALIGNMENT 30

6 Select the residues corresponding to the first peak of the RMSD graph in
the Multiseq program window, by clicking on them. They will be high-
lighted as you click on them (Fig. 26). If you want to select residues in
different regions, hold down the shift key while selecting them.

Figure 27: Highlighted residues shown in bond representation

7 Go back to the VMD Main window. Choose Graphics → Representations...

In the Graphical Representations window, (Fig. 28) you should find that a
new representation has been created for each molecule, with the residues
you have selected as shown in Fig. 26. Look at the representation for
molecule 1fqy and compare the selection with the RMSD graph (Fig. 25)

8 To proceed to the next section, close the Graphical Representations and

the RMSD per Residue windows.
6 EXAMINING THE AQUAPORIN TETRAMER 31

Figure 28: Graphical Representations Window

6 Examining the Aquaporin Tetramer

Aquaporins form homotetramers in cell membranes.The monomers are arranged
side by side in a tight cluster. Each monomer conducts water. A fifth pore
is formed in the center of the tetramer, with a yet unknown function. In the
following we have a closer look at the aqua-glyceroporin GlpF (PDB code: 1lda),
an aquaporin that conducts both water and glycerol.

6.1 Loading Tetramer

In this section we will load the tetramer structure of GlpF and align it with the
other four AQP molecules. A pdb file with the atom coordinates of the tetramer
is provided with the tutorial (see Getting Started).

1 Go to the VMD Main Window

2 Choose File → New Molecule. Another window, the Molecule File Browser,
will appear.

3 Use the Browse button to find the file glpf tetramer.pdb in the directory
aqp tutorial files → PDB, in the tutorial directory.
6 EXAMINING THE AQUAPORIN TETRAMER 32

4 Once you have selected the pdb file, press the Load button in the Molecule
File Browser window to load the molecule. It will take a while for the
tetramer to load

5 The GlpF tetramer is loaded in Lines representation. To have a better

view of the molecule after alignment, you need to change the molecule’s
graphical representation to Tube. From the VMD Main window, choose
Graphics → Representations. In the Graphical Representation window,
change the molecule’s Coloring Method to SegName, and its Drawing
Method to Tube.

Now that you have the tetramer loaded, go back to the Multiseq program win-
dow and select Tools → STAMP structural alignment to align this molecule with
the other four. Note that for this section you have to have the aquaporins 1j4n,
1fqy, 1lda and 1rc2 (Table 1.2) loaded. If you don’t have them, go back and
load the four pdb files, as explained in section 1.

6.2 Examining channel lining and tetrameric contacts

In the previous section, you have identified the conserved residues among the
aquaporin molecules. Here we look where these residues are located in the
tetrameric structure.

Figure 29: Select Residue Dialog

1 Go to the Multiseq program window.

2 Choose Search → Select Residues.

3 The Select Residue dialog will appear (Fig. 29). Select All Sequences.

4 In the following box, select Where Sequence Identity is, ≥, and type in
100 for the value.
6 EXAMINING THE AQUAPORIN TETRAMER 33

5 Click on the Select button. Now you will see the conserved residues high-
lighted in the OpenGL Display window as well as in the Multiseq program
window.

6 Take some time to examine the highlighted parts as shown in Figure 30.
You recognize that most conserved residues line the channel interior. Lo-
cate the conserved residues that are outside the water conducting pore.
Note that these residues are indeed located at the subunit-subunit inter-
faces of the tetrameric GlpF.

Figure 30: GlpF tetramer aligned with other aquaporin molecules. Conserved residues are
highlighted in bond representation (yellow).

7 Before starting the next section, go to the VMD Multiseq program window,
choose the tetramer molecule by clicking on it (It will turn yellow), and
delete the molecule by using the Delete key on your keyboard.
7 PHYLOGENETIC TREE 34

7 Phylogenetic Tree
The Phylogenetic Tree feature, in the Multiseq program, elucidates the structure-
based relationships between the four aquaporins considered.

Structure based phylogenetic trees can be constructed according to the RMSD

or Q values between the molecules after alignment. Here we use the Q values
to obtain the phylogenetic tree of four aquaporins.

The Phylogenetic Tree. A phylogenetic tree is a dendogram, rep-

resenting the succession of biological form by similarity-based clus-
tering. Classical taxonomists use these methods to infer evolution-
ary relationships of multicellular organisms based on morphology.
Molecular evolutionary studies use DNA, RNA, protein sequences or
protein structures to depict the evolutionary relationships of genes
and gene products. In this tutorial we employ QH and RMSD to
depict evolution of protein structure. For a comprehensive expla-
nation of phylogenetic trees, see Inferring Phylogenies by Joseph
Felsenstein5 .

To utilize this feature:

1 Align the structures again, by going to the Multiseq program window and
selecting Tools → Stamp Structural Alignment.

2 In the Stamp Structural Alignment window, select All Structures, and keep
the default values for the rest of the parameters. Press the OK button to
align the structures.

3 In the Multiseq program window choose Tools → Phylogenetic Tree. The

Create Phylogenetic Tree dialog will appear (Fig. 31).

Here you can choose to construct a structure based phylogenetic tree according
to the QH values or RMSD. QH is a variation of Q (see below) that accounts
for both gapped and aligned regions.

What is QH ?. QH is an adaptation of the Q value, and is

essentially a metric for structural homology. QH is comprised of
two terms: Qalign and Qgap. Each time you align two structures,
unless they align perfectly, the structure can be divided into two
parts: the part that aligns (Qalign) and the part that due to insertion
and deletion of the protein sequence, does not align (Qgap). Qgap
accounts for how much the insertion perturbs the core structure
of the protein and for the size of the insertion in sequence and
structure.

5 J. Felsenstein Inferring Phylogenies. Sinauer Associates, Inc.: 2004.

7 PHYLOGENETIC TREE 35

Figure 31: Create Phylogenetic Tree window.

Figure 32: Structure-based phylogenetic tree, for the aquaporins in Table 1.2.

4 Select Structural tree using QH , and press the OK button. A phylogenetic

tree based on the QH values will be calculated and drawn as shown in Fig-
ure 32. Here you can see the relationship between the four aquaporin, e.g.,
how the E.Coli AqpZ (1r2c) is related to human AQP1 (1fqy). Identify
bovine and human aquaporins in the Tree.

5 You can choose to construct the phylogenetic tree of the four aquapor-
ins based on their sequence information, and will be discussed further is
section 8.

6 You can close the Phylogenetic tree window.

7 Quit VMD.
8 EVOLUTIONARY PROFILE OF AQPS 36

8 Evolutionary Profile of AQPs

So far you have learned how to construct a phylogenetic tree using the struc-
tural alignment of AQPs. The sequence alignment can also be used to build a
phylogenetic tree, especially when protein structures are not available. To make
use of both the structural and sequence information, Multiseq now allows you
to merge the two types of alignments and construct a complete evolutionary
profile (EP) for the proteins being studied. In this section, you will learn how
to obtain the EP for AQPs. For more information on using Multiseq program
to perform evolutionary analysis, please refer to the Evolution of Biomolecular
Structure tutorial.

Evolutionary Profile. Due to practical reasons such as the

amenability of the system, protein databases are often biased to-
wards one domain of life, e.g. bacteria. To characterize proteins
based on their patterns in evolution, rather than their occupancy in
the current databases, Multiseq utilizes the QR factorization algo-
rithm to allow for the selection of a minimum non-redundant set of
sequences and structures. The derived non-redundant representa-
tive multiple alignments and statistical representations are termed
evolutionary profiles.

8.1 Configure BLAST for Multiseq

For the following section you will need to install BLAST in your computer.
BLAST is a software that searches through sequence databases and locate
those sequences that are similar to a query sequence. It is available online
at http://www.ncbi.nlm.nih.gov/BLAST/. Here we will install a local copy of
BLAST for Multiseq.

1 Create a directory into which BLAST will be installed.

Examples:
Unix/Linux:/usr/local/blast;
Mac OS X:/Applications/Blast;
Windows: C:\ Blast

2 Extract the archives of BLAST.

Copy the blast installation file for your platform from the aqp-tutorial-files →
blast-install directory to the directory you’ve made. In Unix or Linux, extract
the files by using the command tar zxvf filename. On Mac OS X or Windows,
double-click the file.

3 Do the same for swiss-prot.tar.gz.

Repeat the above two steps: create a directory for swiss-prot, copy the file swiss-
prot.tar.gz from aqp-tutorial-files to the directory you’ve created, and extract it.
8 EVOLUTIONARY PROFILE OF AQPS 37

Figure 33: Choose the directory for BLAST.

4 Set the BLAST installation location in Multiseq.

In the Multiseq program window, choose File → Preferences. Click on the

Software button in the new dialog to bring up the software preferences. Click
on the Browse button in the BLAST Installation Directory section and select the
directory into which you installed BLAST (Fig. 33). Note: You may be asked by
Multiseq to update certain databases before you could continue, if so, click Yes
and wait for Multiseq to finish the update. In Linux or Mac OS X, you may have
a directory named blast-2.2.13 in your installation directory. Pick this directory
if you have it.

8.2 Load Structures for AQPs in all three domains of life

AQPs are present in all three domains of life. To build a complete EP for AQPs,
we will first perform a structural alignment for AQPs in all three domains of
life: Eukaryota, Bacteria, and Archaea. In the previous sections, you have seen
the structures of human AQP1 (1fqy) and E .coli AqpZ (1rc2). Here you will
also need AqpM (2f2b) from Archaea to construct the EP. We have provided
these pdb files in the tutorial files for you.

1 Open a new VMD and load the pdb files 1fqy, 1rc2, and 2f2b one by one.

2 Open the Multiseq program by clicking Extensions → Analysis → Multiseq.

3 In the Multiseq program window, keep the protein structures under VMD
Protein Structures and delete all structures under VMD Nucleic
Structures.

If you loaded your structures by giving the pdb code to VMD (not with the
files we prepared for you), you may have more than one structure for each of
the pdb code you entered, i.e., besides 1rc2 A, you may also have 1rc2 B. This
8 EVOLUTIONARY PROFILE OF AQPS 38

Figure 34: Import structures for AQPs in Multiseq.

indicates that in the original pdb file, there are two different structures for the
protein. The difference between them is usually very small, and does not affect
the alignment we are going to perform. Therefore, simply delete 1rc2 B and
keep 1rc2 A.

8.3 Load Sequences for AQPs in all three domains of life

Now that you have the structures of AQPs loaded, we will use BLAST to find
sequences of AQPs in all three domains of life. Each of the three structures
will be used as a query sequence by BLAST, and sequences in the swiss-prot
database will be compared with them, one at a time. Those sequences similar
to our query sequence will be picked by BLAST and loaded in Multiseq.

1 In the Multiseq window, check the box in front of 1fqy. Then click File →
Import Data.

You will find the same window you’ve seen when loading the pdb structures.
This time, choose From BLAST Search under Data Source and select Marked
Sequences (Fig. 34).

2 Click the Browse button after Databases, and go to the directory where
you extracted the file swiss-prot.tar.gz. You should find a direcotry named
swiss-prot. Go into that directory and select the file uniprot sprot.

3 Choose e−20 for E Score and 1 for Iterations and then click OK.

BLAST is now searching the database with 1fqy as a query sequence. This
should take a minute or two. A new window named BLAST Search Results will
open once the search has finished. Note that the swiss-prot database provided
here only contains sequence data for proteins in this session. You cannot rely
on it for other proteins that you want to investigate. Moreover, the database is
not an updated one, so visit the BLAST online databases if you want the latest
8 EVOLUTIONARY PROFILE OF AQPS 39

results. As you may have noticed, 100 sequences have been found using the query
sequence 1fqy. We will only keep those sequences from the Eukaryota domain,
since our query sequence is from Eukaryota. Later we will find sequences in
Bacteria and Archaea using the query sequences 1rc2 and 2f2b, respectively.
This should make our search more accurate.

4 In the BLAST Search Results window, under Domains, unselect the All list
and select Eukaryota. Click Apply Filter.

You will find that only 87 sequences are left (Fig. 35).

5 Click Accept. The sequences will be loaded in Multiseq.

Figure 35: Search result of BLAST.

6 Check the box in front of 1rc2 and uncheck 1fqy in the Multiseq window.

Now you could repeat the above process and find Bacteria sequences using 1rc2
as a query sequence. You should find 28 sequences from Bacteria. Repeat this
process using 2f2b as a query sequence and get 3 sequences for Archaea.
Before we continue, save your Multiseq session by clicking File → Save Session
and save it as aqp.multiseq. You can load the session later by clicking clicking
File → Load Session. There is a saved aqp.multiseq session in the tutorial files,
in case you’d like to check with it.
8 EVOLUTIONARY PROFILE OF AQPS 40

8.4 Align Sequences Using a Structural Profile

In order to analyze the three structures and the 118 sequences of AQPs together,
we need to first align them. What we will do is to first align the structures using
the STAMP structural alignment tool mentioned in section 2, and then we will
use the structural alignment to guide the sequence alignments.

1 Mark the three pdb structures by checking the boxes in front of them.
Make sure that no other sequences are marked.

2 Click Tools → Stamp Structural Alignment and choose to align Marked

Structures and then click OK.

3 Unmark structures and mark all the sequences. Remove gaps in the se-
quences by clicking Edit → Remove Gaps and then select Remove gaps
from: Marked sequences, and Remove these types of gaps: All gaps.

You could select all the sequences at once by clicking on the first sequence, press-
ing the shift button and then clicking on the last sequence. All the sequences
should appear in yellow now, which means they are highlighted. Press the shift
button and check one box in front of any highlighted sequence. All other boxes
for the highlighted sequences should be automatically checked.

4 Click Tools → ClustalW Sequence Alignment.

A new window named ClustalW Alignment Options should appear (Fig 36). As
we are going to align the sequences using the structural alignment, choose Pro-
file/Sequence Alignment instead of Multiple Alignment in the window. Under
Align marked sequences to group, select VMD Protein Structures, and then click
OK. This should take two or three minutes.
Now you have a complete structural based alignment of the AQPs in all
three domains. Try coloring it by sequence identity by clicking View → Coloring
→ Sequence Identity (Fig.37).

8.5 Construct an Evolutionary Profile for AQPs

Although we have obtained the structures and sequences for AQPs in all three
domains and aligned them together, what we have now is not an evolutionary
profile yet. We still need to get rid of the redundancy in these sequences caused
by the biased databases. Multiseq provides a Sequence QR tool which can be used
to select a minimum non-redundant set from the sequences, using a threshold
specified by the user.

1 Mark all the sequences and make sure that the structures are unmarked.
Click Search → Select Non-Redundant Set from the menu.

A new window named Select Non-Redundant Set should show up (Fig. 38). In
this window, choose Select from → Marked Sequences, and choose Using Sequence
QR. Set the Maximum PID to 75 and then click OK.
8 EVOLUTIONARY PROFILE OF AQPS 41

Figure 36: ClustalW Alignment Options window.

Figure 37: ClustalW alignment result using the structural profile

You should find that some of your sequences are highlighted after the pro-
gram stopped calculating. These represent the non-redundant set that Multiseq
selected for you. Group them together by clicking Options → Grouping → From
Selection and enter “NR set” for the new group. This should put all your high-
lighted sequences into a group named “NR set”. This is the evolutionary profile
(EP) for AQPs. You could now create the phylogenetic tree using the EP of
AQPs: simply delete all the sequences except the ones in the NR set and create
a phlogenetic tree as you did in section 7.
8 EVOLUTIONARY PROFILE OF AQPS 42

Figure 38: Select Non-Redundant Set window.

Evolutionary profile provides an “unbiased” view for the evolutionary rela-

tionship of the proteins in investigation. Using EP, scientists have successfully
identified a new subfamily for the protein cysteinyl-tRNA synthetase. For more
details on constructing EP and performing evolutionary analysis, please refer to
the Evolution of Biomolecular Structure tutorial.